An analysis of self-reported racism and BMI Is self-reported weight and racial discrimination associated with BMI and obesity? Gilbert C. Gee, Ph.D. Annie Ro, M.P.H. Amelia Galvin, Ph.D. David T. Takeuchi, Ph.D. November 5, 2007 APHA Conference An analysis of self-reported racism and BMI

Purpose of Study Research Questions: Is self-reported weight and racial discrimination associated with BMI and obesity among a national sample of Asian Americans? Among immigrants, does this relationship vary by years in the U.S.?

Background Stressors negatively affect health Cortisol release increases fat retention in the abdominal area (Rosmond et al, 1998; Marneimi et al, 2002) May increase intake of high-calorie, fatty foods (Dallman et al, 2003; Dallman et al, 2005) Racial discrimination is a chronic stressor

Related Research Stress due to disadvantaged social status may be related to obesity British civil servants (Rosmond et al., 1998) Bullying or emotional abuse (Gunstad et al., 2006) Discrimination may be associated with weight gain Afro-Caribbean women (Tull et al., 1999) Caribbean adolescents (Chambers et al., 2004)

Conceptual Diagram Racial Discrimination (stressor) Physiological effects: metabolic shifts, change in diet Increased BMI, increased obesity

Caveats Potential confounding with weight discrimination Potential interactive effect: discrimination*years in the U.S.

Conceptual Diagram Years in the U.S. Weight Discrimination Racial Discrimination (stressor) Physiological effects: metabolic shifts, change in diet Increased BMI, increased obesity Is self-reported weight and racial discrimination associated with BMI and obesity among a national sample of Asian Americans? Among immigrants, does this relationship vary by years in the U.S.?

Data National Latino and Asian American Study (NLAAS) Nationally representative psychiatric epidemiological study of Asians and Latinos. Variables include: social position, environmental context, psychosocial factors, acculturation, psychiatric illness, mental health service use

Variables Dependent Variables Measurement BMI (Continuous) Weight (lbs)/Height (in.)2 x 703.49 BMI Categories 1. Underweight/normal = 18.5-24.9 2. Overweight= 25-29.9 3. Obese = over 30

Variables Independent Variables Measurement Reasons for discrimination (Categorical) Race (nationality, ethnicity and skin color) Weight Other (height, gender age, sexual orientation, income, education) None (no report of discrimination) Adapted from Jackson & Williams, 1995 Covariates Mental disorders, self-rated physical health, social desirability, employment, immigration generation, duration in U.S., age, gender, per capita household income

Analysis Dependent Independent Interaction Covariates I: Linear Regression BMI Self-reported discrim. Included II: Multinomial Logistic Regression BMI categories III: Linear Regression w interaction terms Discrim. * years in the U.S.

Study Population Variables of Interest Study Population (n=1956), weighted BMI, mean 24.2 (.2) Weight category Underweight/normal 64.9% Overweight 26.9% Obese 9.3% Discrimination None 25.4% Racial 41.7% Weight 0.8% Other 32.1%

Study Population Immigration and Ethnicity Related Characteristics Characteristics of study Population, weighted Immigration Born outside US 76.94% Born US and parents born abroad 13.68% Respondent and parents born in US 9.38% Ethnicities Vietnamese 12.93% Filipino 21.59% Chinese 28.69% Other API 36.79%

Bivariate Analysis Results Discrimination Normal/ Underweight (%) Overweight (%) Obese (%) BMI (SE) None 70.08 25 4.93 23.48 (.30) Racial 60.81 28.89 10.29 24.44 (.17) Weight 24.61 27.86 47.56 31.24 (1.78) Other 64.63 24.90 10.47 24.34 (.26)

Bivariate Analysis Results Discrimination Overweight (%) Obese (%) None 25 4.93 Racial 28.89 10.29 Weight 27.86 47.56 Other 24.90 10.47

Multivariate Analysis Results BMI continuous Beta (SE) Race discrimination .68 (.34) * Weight discrimination .86 (.24) ** Other discrimination .03 (.91) No discrimination Reference * p>= 0.05 ** p>= 0.001 Controlling for mental disorders, self-rated physical health, social desirability, employment, immigration generation, duration in U.S., age, gender, per capita household income

Multivariate Analysis Results BMI categorical Obesity RRR (95% CI) Race discrimination 2.12 (1.04, 4.33) * Weight discrimination 1.52 (1.25, 1.85) ** Other discrimination 1.05 (.98, 1.12) No discrimination Reference * p>= 0.05 ** p>= 0.001 Controlling for mental disorders, self-rated physical health, social desirability, employment, immigration generation, duration in U.S., age, gender, per capita household income

Multivariate Analysis Results Interaction Terms

Multivariate Analysis Results Interaction Terms

Discussion Increasing BMI and the obese BMI categories were be positively correlated with increased reports of racism Interactive effect is an alternative explanation for the increase in BMI for immigrants the longer they are in the U.S. Explore physiological mechanisms and other structural influences on BMI

Acknowledgements NLAAS: NIMH (U01 MH62209, U01MH62207): PI David Takeuchi & Margarita Alegria U of M Population Studies Center Training Grant: NIA PI: Arline Geronimus

Thank You

Study Population Health Related Characteristics Characteristics of study Population, weighted Mental Disorder (12-Month DSM-IV) 9.5% Self-Rated Physical Health 3.5 (0.04) Social Desirability 2.2 (0.1)

Study Population SES and demographic characteristics Characteristics of study Population, weighted Percent Married 65.39% Employed 64.1% Per capita income 39,3000 (1,600) Mean Age 41.32 years Percent Female 52.55% Region West 67.1% Northeast 16.1% South 7.8% Midwest 9.0%